APPLICANT'S ABSTRACT: Alcohol drinking is a major source of mortality and morbidity. The drinking patterns of the elderly and the near-elderly have not received the same amount of attention that problem drinking in younger age groups has received. The higher frequency of adverse events such as onset of illness and loss of relatives and friends that may lead to heavy drinking is greater for persons over age 50 than for younger individuals. Also, such persons may be more vulnerable to the effects of heavy alcohol use for several reasons. Among these are their general health, and the greater number of medications they take. The impacts of adverse life experiences on drinking behavior and the effects of such behavior on life experiences are not adequately understood, particularly among persons aged 50 and over. Much of the research conducted to date has been based on localized and/or cross-sectional data bases. This study uses two panel data bases, the Health and Retirement Study (HRS, 4 Waves, 1992-1998) and the Asset and Health Dynamics of the Oldest Old (AHEAD, 3 waves 1993-98) to examine the relationship between life experiences and drinking behavior and the converse. These data will be supplemented by another panel, the National Long Term Care Survey (NLTCS, 1982-1999) which has been linked with Medicare claims data for the same years to assess mortality and some dimensions of morbidity as well as the impact of excessive alcohol use on Medicare expenditures. The specific aims of this 3 year study are: 1) to determine the patterns of alcohol consumption over the life cycle; 2) assess the effects of excessive alcohol use on mortality, morbidity, Medicare expenditures, job loss, and divorce, recognizing the possible endogeneity of heavy drinking; and 3) to specify and estimate dynamic models of drinking behavior to examine drinking and its effects on job loss or divorce. Our analytic strategy is to start with description, such as of drinking behavior over time, and then specify models of increasing complexity such as for determinants of drinking behavior, and effects of drinking on the outcomes listed above, leading eventually to estimation of dynamic structural models of drinking behavior and effects.